Richard:
> I think that is pushing it a bit far. It could be that the phrase
> would express possession far more often that it expressed
> apposition, so they were specialised to the longer form expressing
> possession and the shorter from expressing apposition.
Erh, since the nominative and genitive would otherwise be identical
to each other, are you suggesting that there may have been a
situation where [X-s Y-s] with the same endings implied that X
possessed Y despite the potential confusion between cases? This
would place greater importance on word order but word order remains
unimportant in IE and was surely unimportant in pre-IE before the
potential merger of these case endings in the thematic paradigm.
>> Further, possession with the locative is supported
>> by Turkish and therefore a fully natural solution.
>
> Could you explain the Turkish construction to me, please.
Simply in the sense that Turkish expresses English "to have"
(indicating a possession) in terms of a verb "to be" and
a locative. "Money is with me" rather than "I have money",
kinda thing. This and many other examples show that locatives
can express possession.
So in my theory, we _expect_ what we in fact see. We see an
endingless relative particle *-yo suggesting an endingless
locative... I mean, it's certainly not the vocative here and
Jens' trying to explain away his need for a nominative *-s
is pointless when we clearly don't need it in the first place.
Think of the above Turkish idea in terms of a clause with
an intended verb that doesn't surface. So while we might
imagine a construction like */wlkWosyo hWo:kWs/ "eyes of the
wolf", underneath the surface the meaning is "the eyes
(being) with the wolf" / "the wolf (having) eyes" / "the
eyes (had) by the wolf".
The semantics are without a single problem and since we DO
see endingless locatives elsewhere in IE whereas we DON'T see
nominative *-s disappearing after vowels (and no, *so is not
a clear example of this at all), the solution is painfully
obvious at this point.
In fact, Jens insistance that *-syo is from earlier **-syoz
based on his conjecture that *so < **soz is evidently
circular and therefore flawed.
Everything is in fact normal with my construction afais. Jens
insists on irregularities that aren't apparent to provide even
more solutions that aren't necessary, which can only lead to a
wildly erroneous view of Pre-IE.
> As the noun phrase cases would then be marked on
> the 'possessor', the composed genitive ought to have nominative
> and oblique forms at the very least - both *-osyo and **-omyo.
Which is a flaw that Jens can't recover from without more and
more and more conjecture. Multiplication of hypotheses is unsound.
As we all know, there is no **-omyo anywhere and therefore even
less basis for believing that the *s in *-syo was a nominative.
It's clearly a _genitive_ that was going to merge with the
nominative in mLIE unless action was taken.
Of course, we might ask why we don't see *-omyo for the same
reason (where *-om is the gen.pl) but this becomes clear now.
Follow me here: The genitive in *-os could not recover from
homophony with the nominative no matter what. It could not
become *-o:s (since this was already the nominative plural,
ie: *-o- + *-es). It absolutely needed another suffix to
disambiguate it from all other *s-endings in the paradigms.
On the other hand, the plural genitive in *-om needed only to
lengthen to *-o:m (ie: *-o- + *-om) to avoid merger with the
accusative singular. None of this can be explained if we hold
on to the mistaken belief that the *s in *-syo was a nominative.
So, *-s-yo can be nothing other than a genitive with an ancient
locative form that would have become *yo in IE proper if it
weren't for its replacement with a new form at the last moment.
> I much prefer the suggestion that the relative pronoun was an
> uninflected particle, as in non-standard English 'as' (e.g. 'It's
> the poor as gets the blame.')
Well, we can state this for *so because it doesn't ever have a
case ending attached to it (only feminine *-x which is a gender
marking, not a case ending). However, we know that *yo- was
inclined so it's rather impossible to support this view. So,
yet again, we have to accept that the endingless *-yo is indicates
an old endingless locative which was later replaced by a more
analogical, synthetic form in a more recent layer of pre-IE.
Perhaps you became entranced by this idea after accepting that
*so was entirely undeclined in the past. However, you have to
understand that earlier MIE *sa is a special marker very different
from the other demonstratives in IE. It was like English "the" and
similarly, English speakers don't use "the" in the exact same way
as we use "this" or "that". The latter two words have a greater
freedom, usable as a replacement for the noun while "the"
absolutely requires an accompanying noun. So *so is ultimately
a definite particle seperate from the demonstrative system
exemplified by proximal *ko- and distal *to- (ironically the same
system as Tyrrhenian *ka and *ta), relative *yo- and interrogative
*kWo-.
= gLeN